Bending losses

[Paul Hill]

Hi,

I am looking to simulate the losses experienced by a wave as it travels around a waveguide as it bends. I only found one topic alluding to such a thing on the Optiwave forums that mentioned OptiBPM for waveguides with a low degree of bending and OptiFDTD otherwise using Observation Areas and monitoring the Power that is lost over a distance.

I have started with a straight channel waveguide for simplicity and found the mode using an input field that is CW, modal, and added 3 XY observation areas with 3D Data components of Ex Ey Hx Hy and all the pointing vectors selected. When the simulation ends and the Analyser appears I can select Tools>Observation Area Analysis but I am not sure what I am looking at or how this helps me determine the losses of the waveguide.

Could somebody please explain it to me in layman’s terms. I have attached project file but the analyzer results are too large.

Thanks,

Paul.

Attachments:

1. test.fdt

[Damian Marek]

Normally you would want to perform a Guassian Modulated CW input simulation, so that you could calculated the loss over the spectrum. In that case, when you open the Tools->Observation Area Analysis there will be an option to inspect the Power Spectrum for each Observation Area. This gives you the detected power of the total simulation time over the detectors area.

A tutorial on the subject can be found at:

Transmission and Reflection Spectrum from Grating Unit Cell

Let me know if this answers your question.

[Paul Hill]

I followed the instructions in the mentioned tutorial but applied to my 3d example, I could see no options for setting the sampling size as it mentions but I think I managed everything else. When coming to plotting the Power Spectrum however even with the “Normalize With:” check box selected the spectrum shows values ranging as high as 4.

Can you suggest why I am not getting values between 0 and 1?

EDIT***: The above was when I used an observation area very close to the Input plane, the next observation area shows .1%. Is it then just an error due to being too close to the Input Plane that gave me this artifact?

[Steve Dods]

If the bend is not too tight and the waveguide weakly guiding, you can get good results with OptiBPM using the Conformal Mapping feature. I attach a ppt that describes the theory behind this approach.

Attachments:

1. Conformal-Map-Theory.pptx

[Paul Hill]

Sorry I was away for a couple of days, I have followed your suggestion Damian and got data I couldn’t really make sense of as it was modulated the power spectrum was a sinusoidal one that gave values above 1 when normalized with Input Field which confused me, I have attached examples of the data before normalization(Modulated1/2). When I swapped from Gaussian Modulated CW to a simple CW (@ 0.55mu) I got spectra that seemed to make more sense but which were very small (like fempto, atto, small). I am not sure if the data makes sense I have attached an example you could perhaps comment on (ObservationArea1/2).

Steve I am interested in trying out the conformal mapping method as well but can’t find any help on the forums or tutorials on how to implement this, are you able to suggest anything?

EDIT: Apparently I cant attach spectra for security reasons but essentially both were gaussian around the center of 0.55 (the wavelength I picked) with the modulated being sinusoidal (i.e. the envelope was gaussian).

[Damian Marek]

Hi Paul,

We have allowed those data types if you want to re-attach them. Also did you end up modifying the previous project file or is that the design that gave you your results?

Regards

[Damian Marek]

The first thing that I found that should be changed is the boundary conditions you have chosen PBC or periodic boundary conditions for every boundary which I don’t think is advisable in most cases since the power will never dissipate and the simulation domain will be filled with interfering waves. In your case the power would flow from the left to right and when it encounters the right boundary it will loop back to the left and enter the simulation domain again.

This could also be the reason why you are getting transmission power of greater than 1, since the power is propagating in a loop and passing the area detector more than once. I would change all the boundaries to absorbing boundary conditions to get an accurate simulation.

[Paul Hill]

Thank you Damian

Attachments:

1. Modulated1.f2d
2. Modulated2.f2d
3. ObservationArea1.f2d
4. ObservationArea2.f2d
5. test2.fdt

[Steve Dods]

The OptiBPM User Reference.pdf has some information about how to use Conformal Mapping in OptiBPM. Look in the chapter on Regions (p 169). More details are found in the Region dialog box tabs on page 184.

[Paul Hill]

I am now getting sensible data using the FDTD since correcting the boundary conditions Damian spotted using a straight waveguide and have chose it as the best answer as it is the closest to the original problem statement.

Thank you Steve for your help also and I still intend to pursue your suggested method in time for comparison.

[Md.Adnan Riaz]

Hi. I have designed some straight waveguides of different and the simulations are also done but i can not simulate the bend waveguide. there is an error that the input field is invalid or something though i have designed in the same way. can anyone help me?

[Author]

Posts